[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: stupid heteroskedastic question
(That was the original poster's title!)
Say I want to do the basic fix for heteroskedasticity
in the linear model
Y = b0 +b1X1 + b2X2 + noise
by dividing through by 1/X1.
How do people report the results ? Do they
report the new coefficients, or transform back
to the old coefficients ?
Don't do this "by hand". Use weights in the estimation. help weights
(and [U] weight, weighted estimation) for the details. Everything
then is in the context of the original variables (R^2, RMSE, etc.)
Kit Baum, Boston College Economics
* For searches and help try: